Polybot— Autonomous AI robots that harvest greenhouse crops.

Polybot builds fully autonomous farming robots that learn dexterous tasks like picking tomatoes end-to-end from human demonstrations, aiming to make diverse, sustainable polyculture farming economical.

Visit website
Founded
2025
Headquarters
Tübingen, Germany
Legal name
Polybot GmbH

Highlights

Notable milestones from Polybot’s journey — funding rounds, launches, and other moments worth knowing.

milestone

Harvesting full rows reliably

By December the system was harvesting entire greenhouse rows reliably — putting Polybot on course for a fully integrated operational pilot in summer 2026 and first system deliveries in early 2027.

Picking gently and dropping into the crate — the loop that has to run row after row to pay off.
Source
team

Spun out as Polybot GmbH

The project was incorporated as a company, formally spinning out of the ELLIS Institute Tübingen to commercialise its first product: a fully autonomous truss-tomato harvester aimed at greenhouses where labour is scarce and expensive.

Source
milestone

First tomato picked fully autonomously

Months after lab work began in January, Polybot's arm picked its first truss tomato on its own in a greenhouse — the robot learns harvesting end-to-end from human demonstrations rather than hand-coded motions, so it adapts fast to new crops.

The manipulator working a truss in the greenhouse — dexterity learned from watching farmers.
Source
funding

€220K SPRIND grant to spin out of the lab

Germany's Federal Agency for Breakthrough Innovation (SPRIND) backed the team with a seven-month validation grant of roughly €220,000 to make the jump from research project to startup, testing the robot on harvesting fine vegetables for real growers.

Round
Grant
Raised
$229K
Led by
SPRIND
Source
team

Born inside Tübingen's AI research cluster

Polybot grew out of Wieland Brendel's lab, drawing on researchers from the Tübingen AI Center, the ELLIS Institute Tübingen, and the Max Planck Institute for Intelligent Systems — betting that learning-based dexterity could finally automate the hardest jobs in farming.

Source